Power supply project

Status
Not open for further replies.

Dr.EM

New Member
Hello, i'm looking at building one of these;

**broken link removed**

I actually don't have a proper power supply around, and am using 12 AA bateries at the moment :shock: . I like the look of this because it has a wide voltage range which I frequently need. I don't need very much current either, so this should be quite substantial. I'll need to build version 2, as those 2k pots are indeed very illusive as dual gang.

Anyhow, i'm wondering about the transformer. There are 2 i'm looking at;

**broken link removed**
(last item, 20-0-20)

**broken link removed**
(50VA 25-0-25)

I see that the transformer requires earthing of the core, presumably to prevent the secondaries going live in the event of a serious failure. What i'm wondering is how to do this on product one. Is it simply a case of earthing that metal bracket? Is that in contact with the iron core directly or not?

I see that the toroidal type has the double insulation safety rating on the primary, which presumably exemplifies it from requiring an earth? But, I also hear that toroids have high inrush current and require a slow blow fuse, is that likely to have any side effects in this application?

I intend to install this into a plastic box, I see no point in making it metal, and I will have all the mains parts completely untouchable. The earthing (or not) is only to prevent the secondaries becoming dangerous in the extremely unlikely event of a transformer failure.

Also, the fuse is suggested as 2A. Should I lower this to my chosen transformers rating (ie, 1A)?

Thanks for any help
 

Looks like product one has just a shield on the core for mechanical mounting to an earth grounded conductor which will also provide some electomagnetic shielding.

Your concern about higher inrush on toroids is unfounded. Your inrush current will depend on your design, not the topology of the magnetics. Don't worry about that. The double insulation is because there is no safety earth connecting ability.

Are you wanting to float your output ground above earth? If not, the usual safety method would be to earth ground your secondary output and put the whole thing in a metal enclosure that is also earth grounded. Any failure on mains to low voltage side will blow the fuse.
 
Toroids have considerably higher surges when you apply power, once you get to large transformers (say for a 250W + 250W power amplifier, or higher) you generally have to add soft-start circuits, or it will take the mains fuse out fairly regularly.

I would certainly recommend an anti-surge fuse (but it's rare to have a fast blow in the mains input of a transformer anyway). Depending how big your toroid is, you may need to increase the size of the plug fuse, plug fuses are only fast blow, and often require a much larger fuse to prevent random failure.

It was common practice not too long ago to always fit a 13A plug fuse in a TV, even though the mains fuse in the set was only 1.6A or 1.25A - but the switch-on surge of the TV would tend to make the fuse fail every few moths for a 5A fuse, or even a 7A.
 
Ok, cool, so I should use the standard chassis type and connect the bracket securely to earth? I don't fancy installing this in a metal box (expensive among other things), I can't help noticing most power supplies are in them though, is there any special reason I should know of? Or is it just so its rugged.

If i'm using the chassis type, which seems to be reccomended here, I should use a slow blow for the 2A rated fuse on the mains in (remember having to use them for an older project), but the plug can have a 3A, or mabye it will need 5A, got loads of them around to try anyhow. Thanks for your help
 

Basically because it's rugged, but also because you then have a metal case you can earth.
 
Hmm, that floating output sounds like it might be useful, to eliminate earth loops etc. If I went with the toroidal type transformer then, 60va, 25-0-25v, I will need to use a slow blow fuse on the mains input, rated at 2A according to the schematic, although the transformer is 1.2A? By what you've said, the plug may well need to use a 5A, but i'll just try a 3A and see how it lasts.

Also, the rating for those 4700uf caps, is 35v suitable? I imagine it would be as only 25v is across each one, but i'm not totally sure about it? I would probably use 2x 2200uf ones anyhow, to save space mainly.
 

The 1.2A is the secondary current, the primary will only be about 300mA, this is a pretty small toroid, and you shouldn't have any surge problems.

Also, the rating for those 4700uf caps, is 35v suitable? I imagine it would be as only 25v is across each one, but i'm not totally sure about it? I would probably use 2x 2200uf ones anyhow, to save space mainly.
Click to expand...

35V is probably too low, I would expect about 40V out of the bridge rectifier without a load - go for 63V caps. The 25V is the AC voltage out of the transformer under full load, when rectified it will be 1.414 times that, minus the voltage drop across the bridge. But without a load the voltage will be higher, and will almost certainly exceed 35V! - capacitors aren't something you can push very much, if it says 35V volts it doesn't mean 36V will be OK.
 
Ah ok, 63v it'll have to be then, perhaps a few smaller ones.

What do you think of the fuse then? Should it just be 2A like the diagram says, or more the order of 300ma like you suggested, or does it really not matter that much?
 
Dr.EM said:
Ah ok, 63v it'll have to be then, perhaps a few smaller ones.

What do you think of the fuse then? Should it just be 2A like the diagram says, or more the order of 300ma like you suggested, or does it really not matter that much?
Click to expand...

2A would be fine, even though it's a fairly small transformer there will still be a reasonably high surge when it powers up. A 300mA fuse would probably blow almost everytime you turned it on.
 
Ok, thanks for all your help, got to order the parts now, looks like it might be more than I thought :shock:
 
Got the parts and i'm putting them together now. I notice that one of the regulators, the LM 337 -v one, looks very different to the 317. Rapid says that they are both NSC devices, but they look very dissimilar. I checked the unusual looking 337, and the pinout is correct. It says Fm E28 across the top if that means anything to anyone :? . I'm just wondering if its going to work in harmony with the other (ie, when one is set to +9v, with the other produce exactly -9v?). Any ideas?
 

Don't worry about it, in any case both are ADJUSTABLE regulators!.
 
Ok, i've fixed it all to stripboard, and i've tried testing it with what is probably a very innapropriate supply, +8.4/-4.7v, taken from some apparently quite drained batteries. Either way, i'm not convinced this is working as it should. I can't see any errors on the board but here's the readings i'm getting from it:

With the control set clockwise:
+1.52/-1.50v

With the control set anticlockwise:
+1.94/-1.92v

Thats wierd for a start as it should be higher with the control set clockwise. I've triple checked that its connected as specified. But other than that, these seem to be very small voltage readings, even for that supply. The voltages at the bases of the transistors go from 0.33-0.68v, also very small. No part is getting hot, but there is no load connected. I've used the BC549 and BC559's as specified, although mine are in fact BC549C and BC559C if that matters, I thought that the C meant it was a higher voltage rating? At the moment, i'm suspicious of those, although this may be normal operation for such a low voltage/imbalanced supply :?
Any help much appreciated
 
Hi Dr. EM,
Your batteries are dead and can't supply the 12.5mA required by the 100 ohm resistors. The minimum input voltages for the ICs are about 3.3V plus 0.6 for the transistors. So with a 3.9V input the circuit's output voltage can't be more than about 1.9V which is its minimum output voltage with any higher input voltage. Try each section separately with wall-wart DC supply that has enough volts to alow the circuit to work.

If the pots operate backwards with a higher input voltage then they are wired wrong. When the slider shorts the base to the emitter of a transistor then the circuit's output voltage should be at its highest, which is about 2V less than its input voltage. :lol:
 
Thanks for your reply
I've tried since wiring all those batteries together in series to give me 14.3v, and using the positive regulator only, but then it gave me the voltage 1.7 continously, with no change regardless of the the pots position :? . Now thats really wierd, but i'll try and find a wall wart and use that, think I may have a 12v one lying around. If it still won't work, would it be a good idea to change those transistors for the trusty 3904 and 3906s :lol: ? I honestly can't see any errors in my layout, all I can think if it still won't work are overheated transistors or that the original design is somehow incorrect :shock:
 
Dr.EM said:
Thanks for your reply
I've tried since wiring all those batteries together in series to give me 14.3v
Click to expand...

Wiring dead batteries in series doesn't stop them being dead!. You need to get a input voltage to the circuit, you can't do that with flat batteries.
 
I don't think you're gonna like the voltage vs pot position in the version with the 20k pots and transistors. With 2k pots, you get a linear rotation angle-to-voltage relationship. With the transistors, the curve looks exponential, with most of the change happening near the clockwise end of the rotation (clockwise --> higher voltage). I ran a sim that shows this (see below).
Phi (horizontal axis) is rotation angle as a decimal fraction of full scale.
 

Attachments

  • ps1_waves.gif
    5.7 KB · Views: 999
  • ps1_sch.gif
    4.7 KB · Views: 1,005
Hi Ron,
You and your sim are correct, the project is grossly flawed. :lol:
I didn't yet but will look on the author's forum to see any discussion about it.

I figure that the circuit will work the same if a common-emitter NPN transistor was used in the LM317 circuit, and visa-versa.
When the pot is halfway, the transistor will have 1.45V across it and the output voltage from the regulator will be that 1.45V plus the 1.25V reference which equals only 2.7V.

I think this circuit will work pretty well:
 

Attachments

  • lm317_voltage_adjust.png
    8.5 KB · Views: 974
Ok, i'll definately try this with a 12v wall wart if I can dig it out

Ron H, I completely believe you, I had an idea this is how it would work. To be honest, I don't think the version with the transistors has ever been fully tested, or at all. If it does turn out very much like that, then I may have to try and get 2 seperate 2k/2.5k pots, and adjust both voltages seperately.

Or, what do you think to using a 1k with a 470ohm at each end or an arrangement like that? I'm assuming that would make it a smaller range of voltages, but i'm not likely to need below +/-9v or above +/-18v. Would that still be linear? If it would, i'll probably grab a 1k dual gang from Maplin and use it like that, removing the transistor and 2.2k as required.
 
Wierd post at the same time there :lol:

I tried this with the wall wart, which actually reads 17vdc.

With the board I get from 1.90v to 2.13v, with the 2.13v being when the control is set more or less central :? past central, it drops back to 1.98 or so.
Totally useless if nothing else :lol:
 
Status
Not open for further replies.
Menu
Log in
Register
Cookies are required to use this site. You must accept them to continue using the site. Learn more…